Analysis Method of Electrical Energy Substitution Potential Based on Time Series and BP Neural Network

China's energy consumption faces problems such as low development efficiency and large fossil energy reserves, and the energy structure needs to be optimized. Electric energy substitution can use electric energy to replace loose coal and oil in the energy consumption terminal, which can effectively promote the optimization and adjustment of China's energy structure and has a good development prospect. In view of this situation, a prediction method combining time series and BP neural network is proposed. The time series model based on three-index smoothing method is predicted by electric alternation, and the prediction result is modified by BP neural network. Energy use data from Shanghai Municipal Bureau of Statistics were used as forecast and comparative analysis data. The calculation results show that compared with the unilateral prediction method, the method combining time series and BP neural network can significantly improve the prediction accuracy of electric energy substitution, and the prediction model can also accurately predict the electric energy substitution potential of Shanghai from 2022 to 2025, which provides a certain reference for Shanghai to make energy substitution plans.